Deep Data Hiding for ICAO-Compliant Face Images: A Survey

View on arXiv ← Back to list

Authors: Jefferson David Rodriguez Chivata, Davide Ghiani, Simone Maurizio La Cava, Marco Micheletto, Giulia Orrù, Federico Lama, Gian Luca Marcialis

Published: 2025-08-26 14:41:12+00:00

Comment: In 2025 IEEE International Joint Conference on Biometrics (IJCB)

AI Summary

This survey investigates digital watermarking and steganography, particularly deep learning-based methods, as complementary solutions for persistent integrity verification of ICAO-compliant facial images. It addresses the limitations of traditional Presentation Attack Detection by enabling post-capture protection against threats like morphing and deepfakes. The paper provides a comprehensive analysis of state-of-the-art techniques, evaluating their potential and drawbacks for secure deployment in real-world identity systems.

Abstract

ICAO-compliant facial images, initially designed for secure biometric passports, are increasingly becoming central to identity verification in a wide range of application contexts, including border control, digital travel credentials, and financial services. While their standardization enables global interoperability, it also facilitates practices such as morphing and deepfakes, which can be exploited for harmful purposes like identity theft and illegal sharing of identity documents. Traditional countermeasures like Presentation Attack Detection (PAD) are limited to real-time capture and offer no post-capture protection. This survey paper investigates digital watermarking and steganography as complementary solutions that embed tamper-evident signals directly into the image, enabling persistent verification without compromising ICAO compliance. We provide the first comprehensive analysis of state-of-the-art techniques to evaluate the potential and drawbacks of the underlying approaches concerning the applications involving ICAO-compliant images and their suitability under standard constraints. We highlight key trade-offs, offering guidance for secure deployment in real-world identity systems.

Key findings

The survey concludes that deep learning-based fragile and semi-fragile data hiding methods, particularly those utilizing Invertible Neural Networks (INNs) and Transformers, are the most promising for ICAO-compliant facial image certification. These methods demonstrate a strong balance between imperceptibility, selective robustness (resisting benign transformations but failing under malicious ones), and reliable blind extraction. Only a subset of existing techniques meets the combined requirements for secure deployment, highlighting the need for specialized strategies tailored to biometric identity constraints.

Approach

The paper surveys digital watermarking and steganography techniques, focusing on deep learning-based fragile and semi-fragile methods, to embed tamper-evident signals directly into ICAO-compliant face images. It analyzes these approaches based on their imperceptibility, robustness, capacity, and security to determine their suitability for persistent integrity verification and traceability against manipulations like deepfakes and morphing.

Datasets

DIV2K, COCO, LFW, PASCALVOC, ImageNet, CelebA, Paris Street, MSCOCO

Model(s)

UNKNOWN

Author countries

Italy

← Previous